Nvidia's Grip on China's AI Chip Market Slips to 34% as Huawei Closes In

For more than a decade, Nvidia's dominance in AI accelerators looked unassailable. Its CUDA software stack, its hardware cadence, and its grip on hyperscaler budgets made it the default choice for anyone training or serving large models. But new market data out this week suggests that in China — once projected to be one of Nvidia's largest growth markets — that dominance is now eroding faster than almost anyone expected.

According to an IDC report seen by Reuters on October 13, Nvidia's share of China's data-center AI chip market fell to 34% in the first half of 2026. The company shipped 678,000 AI accelerators in mainland China between January and June, a 44% decline compared to the same period a year earlier. Meanwhile, Huawei shipped 572,000 accelerators, nearly doubling from just over 300,000 units a year earlier, capturing 28% of the market and narrowing the gap with Nvidia to just six percentage points.

Export Controls Bite, Domestic Chips Surge

Two forces are driving the shift. The first is Washington's escalating export control regime, which has progressively cut off Nvidia's most advanced data-center GPUs from the Chinese market. The second is Beijing's aggressive, well-funded push for semiconductor self-sufficiency, which has turned domestic AI silicon from a science project into a credible commercial alternative.

Huawei has been the biggest beneficiary. In September, the company unveiled more than ten new chipsets spanning AI accelerators, processors, and connectivity components — a product line aimed squarely at the infrastructure territory long dominated by Nvidia, AMD, and Intel. Huawei has also invested in software ecosystem play, working with partners to offer alternatives to Nvidia's CUDA moat.

And Huawei is not alone. Of the eleven Chinese semiconductor suppliers IDC tracks, nine increased their year-over-year shipment volumes. Alibaba's chip-design unit T-Head held 7% of the market, followed by Cambricon at 5%, Baidu's Kunlunxin at 4%, and Hygon at 3%. That is a broad-based domestic ramp, not a single-company story.

A Structural Shift, Not Just a Bad Quarter

Perhaps the most striking finding in the IDC data is what it reveals about the shape of the market itself. Total AI accelerator shipments in China reached 2.008 million units in the first half of 2026 — the market as a whole is still growing fast. IDC projects total shipments will expand 19% for the full year.

But the composition of that growth is changing. GPU shipments actually fell 26% to 991,000 units, while shipments of non-GPU AI accelerators — chips specialized for inference and specific model workloads — surged nearly 80% to 1.016 million units. For the first time, non-GPU accelerators outsold GPUs by total volume in China. That crossover matters: it suggests Chinese buyers are increasingly designing their AI infrastructure around purpose-built domestic silicon rather than trying to substitute for unavailable Nvidia GPUs.

IDC did not specify which Nvidia products made up its 34% share, or whether the figure includes restricted chips such as the B200 and B300, which have reportedly been smuggled into the country despite export bans. Nvidia did not respond to requests for comment on the report.

Why It Matters Beyond China

The implications of this data ripple well beyond one national market.

First, it suggests that export controls are producing exactly the outcome Beijing wanted: a forced acceleration of domestic chip development. Each round of restrictions has strengthened the business case for Chinese vendors, and the shipment numbers now show that the policy is working — just not in the direction Washington may have intended.

Second, China's AI compute build-out is not slowing down. Combined capital expenditures from Alibaba, ByteDance, and Tencent has more than doubled to roughly RMB 700 billion (about $104 billion), and reports indicate Chinese firms have bought more than a million Nvidia gaming and workstation chips this year to plug compute shortfalls. Demand is insatiable; the supply is simply being redirected.

Third, the software moat is being tested. Nvidia's ecosystem advantage has always been CUDA, but the rise of inference-optimized accelerators and open-weight models increasingly tuned for domestic hardware erodes the lock-in that once looked permanent.

What to Watch

The second half of 2026 will answer three questions. Does Nvidia's share keep falling, or stabilize as it ships whatever compliant products it can? Does Huawei close the remaining six-point gap and potentially take the lead in its home market? And do non-GPU accelerators extend their volume lead over GPUs, confirming a structural transformation in how AI compute is built?

One thing is clear: the era in which Nvidia could treat China as home turf is over. What was once a borderless market for American AI silicon has become a contested frontier — and for the first time, the domestic challengers have the shipment numbers to prove they belong in the fight.